The maxispan procedure makes the phonological similarity effect disappear while increasing recall performance.

Based on the hypothesis of two maintenance mechanisms of verbal information in working memory, an articulatory loop and an attentional executive loop, Barrouillet et al. predicted and observed that facilitating the optimal use and separation of these two systems results in a strong increase in recall performance. They developed for this purpose the maxispan procedure, in which participants cumulatively rehearse aloud a limited number of the first items of the series (i.e., three or four) and keep rehearsing them until the end of the series before recall. Beyond increasing recall performance, the model also predicts that the maxispan procedure should also abolish the phonological similarity effect (PSE, the poorer recall of phonologically similar than dissimilar items) in both the rehearsed and the nonrehearsed items by permitting the perfect maintenance of the former in a nonoverloaded articulatory loop and preventing storage of phonological traces of the latter in the attentional system. However, the PSE should reappear if too many items are verbally rehearsed in the maxispan procedure. In this case, the overload of the articulatory loop should lead to offload its content into the attentional system where phonologically similar traces are prone to confusion. We tested and verified these hypotheses in two experiments.